From: Sheng Yong <shengyong1@xiaomi.com>

When committing new super block, bio is allocated but not freed, and
kmemleak complains:

  unreferenced object 0xffff88801d185600 (size 192):
    comm "kworker/3:2", pid 128, jiffies 4298624992
    hex dump (first 32 bytes):
      00 00 00 00 00 00 00 00 80 67 c3 00 81 88 ff ff  .........g......
      01 08 06 00 00 00 00 00 00 00 00 00 01 00 00 00  ................
    backtrace (crc 650ecdb1):
      kmem_cache_alloc_noprof+0x3a9/0x460
      mempool_alloc_noprof+0x12f/0x310
      bio_alloc_bioset+0x1e2/0x7e0
      __f2fs_commit_super+0xe0/0x370
      f2fs_commit_super+0x4ed/0x8c0
      f2fs_record_error_work+0xc7/0x190
      process_one_work+0x7db/0x1970
      worker_thread+0x518/0xea0
      kthread+0x359/0x690
      ret_from_fork+0x34/0x70
      ret_from_fork_asm+0x1a/0x30

The issue can be reproduced by:

  mount /dev/vda /mnt
  i=0
  while :; do
      echo '[h]abc' > /sys/fs/f2fs/vda/extension_list
      echo '[h]!abc' > /sys/fs/f2fs/vda/extension_list
      echo scan > /sys/kernel/debug/kmemleak
      dmesg | grep "new suspected memory leaks"
      [ $? -eq 0 ] && break
      i=$((i + 1))
      echo "$i"
  done
  umount /mnt

Fixes: 5bcde4557862 ("f2fs: get rid of buffer_head use")
Signed-off-by: Sheng Yong <shengyong1@xiaomi.com>
---
 fs/f2fs/super.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c
index ebda6834ac2c..57adeff5ef25 100644
--- a/fs/f2fs/super.c
+++ b/fs/f2fs/super.c
@@ -3701,6 +3701,7 @@ static int __f2fs_commit_super(struct f2fs_sb_info *sbi, struct folio *folio,
 		f2fs_bug_on(sbi, 1);
 
 	ret = submit_bio_wait(bio);
+	bio_put(bio);
 	folio_end_writeback(folio);
 
 	return ret;
-- 
2.43.0

On 2025/6/7 14:41, Sheng Yong wrote:
> From: Sheng Yong <shengyong1@xiaomi.com>
> 
> When committing new super block, bio is allocated but not freed, and
> kmemleak complains:
> 
>    unreferenced object 0xffff88801d185600 (size 192):
>      comm "kworker/3:2", pid 128, jiffies 4298624992
>      hex dump (first 32 bytes):
>        00 00 00 00 00 00 00 00 80 67 c3 00 81 88 ff ff  .........g......
>        01 08 06 00 00 00 00 00 00 00 00 00 01 00 00 00  ................
>      backtrace (crc 650ecdb1):
>        kmem_cache_alloc_noprof+0x3a9/0x460
>        mempool_alloc_noprof+0x12f/0x310
>        bio_alloc_bioset+0x1e2/0x7e0
>        __f2fs_commit_super+0xe0/0x370
>        f2fs_commit_super+0x4ed/0x8c0
>        f2fs_record_error_work+0xc7/0x190
>        process_one_work+0x7db/0x1970
>        worker_thread+0x518/0xea0
>        kthread+0x359/0x690
>        ret_from_fork+0x34/0x70
>        ret_from_fork_asm+0x1a/0x30
> 
> The issue can be reproduced by:
> 
>    mount /dev/vda /mnt
>    i=0
>    while :; do
>        echo '[h]abc' > /sys/fs/f2fs/vda/extension_list
>        echo '[h]!abc' > /sys/fs/f2fs/vda/extension_list
>        echo scan > /sys/kernel/debug/kmemleak
>        dmesg | grep "new suspected memory leaks"
>        [ $? -eq 0 ] && break
>        i=$((i + 1))
>        echo "$i"
>    done
>    umount /mnt
> 
> Fixes: 5bcde4557862 ("f2fs: get rid of buffer_head use")
> Signed-off-by: Sheng Yong <shengyong1@xiaomi.com>

Reviewed-by: Chao Yu <chao@kernel.org>

Thanks,